立命館大学グローバルCOE　GCOEセミナー

2009年11月17日

第６６回　GCOEセミナー

1. ワラワット チョエンサワット Worawat Choensawat（RA / デジタルアーカイブ技術研究班 / 立命館大学大学院理工学研究科 D1）
"Segmentation of Motion Capture Data Using Neural Networks"

2. キンカーン スックハナピバーン Kingkarn Sookhanaphibarn（PD / デジタルアーカイブ技術研究班）
"Analysis of visitors' behaviors in 3D virtual museums"

3. アレハンドロ トレド ノラスコ Alejandro Toledo Nolasco（RA / デジタルアーカイブ技術研究班 / 立命館大学大学院理工学研究科 D1）
"Similarity and Dissimilarity Patterns in Name Occurrences of Japanese Historical Diary"

4．崔 雄 Woong Choi（PD / デジタルアーカイブ技術研究班）
"Character Animation of Performing Arts in Second Life"

日時：2009年11月17日（火）　18:00－19:30（6限）
場所：【衣笠】立命館大学アート・リサーチセンター　多目的ルーム
 【BKC】インターネット（Power Live）をご利用ください。
参加無料(予約不要）
 
※一般の方もインターネットでセミナーにご参加いただけます。
※インターネットでの参加を希望される方は、kkt27007■lt.ritsumei.ac.jp（■を@に置き換えてください）まで、その旨ご連絡ください（担当、楠井）。
※本拠点研究メンバーですでにID、PWをお持ちの方は連絡は不要です。
※配布資料は開催当日午後より下記のURLからご覧いただけます（期間限定）。

https://www.arc.ritsumei.ac.jp/dhjac/ppt2009/haihusiryo-index.html

発表要旨は、「続きを読む」をご覧下さい。

1. ワラワット チョエンサワット Worawat Choensawat（RA / デジタルアーカイブ技術研究班 / 立命館大学大学院理工学研究科 D1）
"Segmentation of Motion Capture Data Using Neural Networks"

For digital archiving and quantitative body motion analysis of traditional dance and performing arts, human motion data obtained by using the motion capture systems has been essential processing information. Segmentation of the body movement is an important pre-processing task for performing various quantitative and/or qualitative researches on dance body motion.  This paper aims at segmentation of human motion data using neural networks (NN). The NN model applied in this paper is a basic feed-forward network with a supervised learning approach. The hypothesis of this work is to discover the segmentation frame based on the velocity of body movement and the joint angles. In the experiment, the numbers of playback frames that can yield the best accuracy are investigated.

2. キンカーン スックハナピバーン Kingkarn Sookhanaphibarn（PD / デジタルアーカイブ技術研究班）
"Analysis of visitors' behaviors in 3D virtual museums"

With advanced technologies in 3D computer graphics and the Internet, people can visit 3D virtual museums via the Internet, where the multi-user virtual environments (MUVE) are provided. For some popular museums in Second Life (SL), its residents wrote comments that there is no significant discernable difference between their experiences in the physical museums and those in the virtual museums. The patterns of visitors’ behaviors in art galleries and museums have been first investigated and reported by Veron and Levasseur (1983) where the categories of visiting styles can be explained by a metaphor of four animals: ant, fish, grasshopper, and butterfly. The ant visitors spend a lot of time observing almost all the exhibits. The fish visitors most of the time move around in the centre of the room and usually avoid looking at exhibits' details. The butterfly visitors stop frequently to look for more information but their stopover times are varying. Last, the grasshopper visitors who have a  set of specific preferences for some pre-selected exhibits spend a lot of time observing them while ignore the others.

3. アレハンドロ トレド ノラスコ Alejandro Toledo Nolasco（RA / デジタルアーカイブ技術研究班 / 立命館大学大学院理工学研究科 D1）
"Similarity and Dissimilarity Patterns in Name Occurrences of Japanese Historical Diary"

Diaries provide historians and people an insight into the past. As historical documents, diaries are important because they are written contemporaneously with the events they record, serve as great educational tools, and at least from the author's perspective, reveal true stories. This presentation describes an extension of our previous system (Toledo, Thawonmas, Maeda, & Kimura, 2009), which uses a stacked graph to dynamically analyze the time series of aristocrat names mentioned in the Japanese Hyohanki diary. To better exploit the system, we propose an improvement on the interaction control with a view to obtain patterns of similarity and dissimilarity in the name occurrences of those names. We measure the similarity distances between the time series using the Euclidean method and found some patterns that can be well used by experts in history to make correlations with historical artifacts.
 

4．崔 雄 Woong Choi（PD / デジタルアーカイブ技術研究班）
"Character Animation of Performing Arts in Second Life"

In this research, we set ourselves the task to record and collate Japanese Traditional Performing Arts, motion data and audio recordings and develop a pipeline method to create a real-time animation combining these within their context within Second LifeTM (SL). For this we specifically focused on assets from Utaura of Noh in Japanese Traditional Performing Arts. We propose a method that synchronously controls motion, sound and subtitle of a performance and transmits this information to avatar of audience in SL. In our research, we successfully collated intangible cultural material through digital capture methods and migrated them though the pipeline method into the SL environment to produce an integrated real-time version of Utaura within the context of the Noh theatre.
 

発表インフォメーション | by PD | 記事詳細 | コメント[0] | トラックバック[0] [編集]

コメントする